A linear-algebra approach that represents multiplication by coordinate functions on the quotient algebra R/I (where I is an ideal) as finite matrices relative to a chosen basis; eigenstructure of these multiplication matrices encodes the coordinates and multiplicities of the common zeros of the ideal when the quotient is finite-dimensional.

Principle

Pick a basis of the quotient algebra R/I (often monomials modulo I); construct the linear map 'multiply by x_j' for each coordinate x_j and represent it as a matrix in that basis. The joint eigenvalues of these commuting multiplication matrices correspond to the coordinates of isolated solutions, and eigenvectors/eigenspaces capture evaluation functionals and multiplicity structure.

Demonstration

For a zero-dimensional ideal in k[x,y], choose a monomial basis {b1,...,bn} of k[x,y]/I. Compute the matrix M_x of multiplication by x: M_x * [c]_B = [x·c]_B. Diagonalizing (or simultaneously triangularizing) M_x and M_y yields eigenpairs whose eigenvalues give x- and y-coordinates of solutions; if an eigenvalue has algebraic multiplicity >1 and the Jordan structure is nontrivial, it signals multiplicity. Numerically this underpins eigenvalue-based solvers that extract roots from linear algebra computations on quotient bases.

Misapplication

Applying the method when the quotient ring is infinite-dimensional (positive-dimensional variety) or using a basis that does not span the quotient leads to invalid matrices; poor conditioning or ignoring that multiplication matrices may not be diagonalizable in the presence of multiplicities can produce misleading numerical eigenvalues and incorrect root multiplicities.

Consequence

When applicable, the multiplication matrix method reduces solving a polynomial system to numerically robust linear algebra operations (matrix construction, eigen-decomposition), providing direct access to solution coordinates and multiplicities, and interfaces well with numeric linear algebra libraries; but it depends on a correct finite-dimensional quotient basis and can be sensitive to numerical conditioning and basis choice.

Reversal

Elimination/resultant-based approaches or homotopy continuation: instead of constructing quotient multiplication operators, eliminate variables to produce univariate polynomials or continuously deform systems to track solutions. These methods avoid forming a quotient basis but trade off other complexities like degree blow-up or path-following costs.

Boundary

Valid primarily for zero-dimensional ideals (finite number of solutions) over algebraically closed fields or with extension fields; requires an explicit finite basis of R/I (Gröbner or border basis) and attention to numerical conditioning, field of coefficients, and possible need for deflation techniques to handle nontrivial multiplicities.

Semantic Tension

Tension exists between the multiplication-matrix viewpoint (algebraic representation via quotient operators) and elimination approaches (symbolic elimination or resultants): both recover roots but differ in data structures, numerical stability, and how multiplicity information is represented (eigenstructure vs multiplicity of factors).

Synthesis

The multiplication matrix method realizes polynomial root finding as an eigenproblem by representing coordinate multiplication on the finite-dimensional quotient algebra as matrices: solved eigenpairs reveal solution coordinates and multiplicities, provided one has a correct finite basis and manages numerical conditioning appropriately.